identification of the latest version (not the latest REC)

[molinaro-m]

In VOUnits, under the discussion for issue #38, the proposed solution for the issue is to:

• not use an Erratum to the existing REC
• fix the text to solve the issue
• wait for a future update of the REC to have the change applied

(sorry @msdemlei for the personal summary).

The one above is one example of an uncovered situation in the document process.

Errata already are not merged within the existing REC document. You have to go to the landing page of the document, see there's an Erratum and read it.

The solution above makes it even less visible the change in the specification (editorial fix in the case above). You have to follow the "draft" release in github to understand what has been fixed and in no place on the repository this is visible nor alerted.

Should we consider, in DocStd, a solution for the latest update of a document that includes merging both the above cases (and maybe others I don't see now)?

If so: how would the user of the document repository know that something has happened to the document s/he thought s/he had already read before? (I do remember resistance in using micro-versioning)

[msdemlei]

On Fri, Nov 28, 2025 at 03:29:05AM -0800, Marco Molinaro wrote: molinaro-m created an issue (ivoa-std/DocStd#9) Errata already are not merged within the existing REC document. You have to go to the landing page of the document, see there's an Erratum and read it.

I think we'd like to change that at least for Documents on github. There are a few technical challenges (in particular: we probably want to record the hash of ivoatex the REC was built with), but nothing really serious I can see. But that's a bit tangential.

The solution above makes it even less visible the change in the specification (editorial fix in the case above). You have to follow the "draft" release in github to understand what has been fixed and in no place on the repository this is visible nor alerted.

I'd say this doesn't regulation; if this were more than an editorial fix that ought to go into the next release and should not be forgotten until then, I'd say it ought to be an erratum. However, a slight inconsitency in citation really is something between editors and authors in my book and nothing the readers of a Spec need to be concerned with. Hence, from me,

Should we consider, in DocStd, a solution for the latest update of a document that includes merging both the above cases (and maybe others I don't see now)?

is a clear no.